MSCT轴向分辨力和图像噪声影响因素分析

目的:探讨多排螺旋CT（MSCT）轴向分辨力或断层灵敏度曲线（SSP）和图像噪声的影响因素。方法:使用临床常用的腹部扫描模式，采用不同直径的模板，重建层厚、螺距、电压（kV）、重建算法等扫面参数，进行MSCT扫描，对不同参数对断层图像SSP和图像噪声的影响进行统计学分析。结果:当螺距和准直器宽度保持不变，不同层厚和重建算法得到的SSP的半高宽FWHM值基本保持不变（P>0.05）；当重建算法和准直器宽度保持不变，不同螺距和层厚得到的SSP的半高宽FWHM值基本保持不变（P>0.05）；不同准直器宽度，同螺距和层厚得到的SSP 的半高宽FWHM值基本保持不变（P>0.05）；随着层厚、mAs增加，图像噪声减小（P<0.05）；随着kV增加，图像噪声随之减小，不同重建算法下图像噪声存在明显差异（P<0.05）。结论:卷积重建算法、螺距和准直器宽度对SSP的影响很小，螺距对图像噪声影响很小，而层厚、重建算法、mAs、kV对图像噪声影响大。层厚、mAs、kV增大，图像噪声减小；重建算法分辨率越高，图像噪声越大。

Analysis of factors affecting MSCT axial resolution and image noise

Objective To explore the factors affecting multi-slice spiral CT (MSCT) axial resolution or section sensitivityprofile (SSP) and image noise. Methods With the commonly-used abdominal scan mode, the templates of different diametersand different parameters, such as reconstruction layer thickness, pitch, voltage and reconstruction algorithm, were adopted forMSCT scanning, thereby statistically analyzing the effects of different parameters on SSP of cross-sectional images andimage noise. Results When pitch and collimator width remained unchanged, the full-width at half-maximum (FWHM) ofSSP obtained by different layer thicknesses and reconstruction algorithms remain basically the same (P>0.05). Whenreconstruction algorithm and collimator width remain unchanged, the FWHM of SSP obtained by different pitches and layerthicknesses also remain basically the same (P>0.05). With different collimator widths, the same pitch and the same layerthickness, the FWHM of SSP remained basically the same (P>0.05). As the layer thickness and mAs increased, the imagenoise decreased (P<0.05). With the increase of kV, image noise was decreased. There were significant differences in imagenoise when different reconstruction algorithms were adopted (P<0.05). Conclusion Convolution reconstruction algorithm,pitch and collimator width have little effect on SSP, and pitch has trivial effect on image noise, while layer thickness,reconstruction algorithm, mAs and kV have significant effects on image noise.With the increases of layer thickness, mAs andkV, the image noise is decreased. The higher the resolution of the reconstruction algorithm is, the larger the image noise is.